A Dynamic Hierarchical Genetic-Fuzzy Sugeno Network

A class of predictive algorithms that use genetic-fuzzy trees has risen to prominence in recent years. These algorithms are defined by three features: fuzzy logic, cascading decision trees, and genetic optimization. This paper presents an algorithm in this class that uses a cascading tree of Takagi-Sugeno systems, where the structure of the tree and the order of the elements within the tree are calculated procedurally by genetic algorithm. The resulting structure is unlike a tree and more akin to a honeycomb. This algorithm is shown herein to compare favorably with similar classes of predictive algorithms.